Background: To date, the difference in microRNA (miR) expression profiles in tears of dry eye patients and healthy people has not been reported. In current study, we evaluated the significance of miRs and transforming growth factor beta2 (TGF-β2) in distinguishing dry eye. Methods: 138 patients with dry eye from October 2017 to October 2018 were selected. During the same period, 138 healthy persons were collected. All patients were followed up for 12 months through outpatient, telephone or medical records, and the time of corneal injury was recorded. Results: Compared with healthy people, TGF-β2 levels in dry eye patients were significantly decreased (P<0.05). Array analysis, predictive software, and dual-luciferase reporter assays showed that miR-450b-5p, miR-1283, and miR-3671 can target TGF-β2 expression. Tear miR-450b-5p, miR-1283, and miR-3671 levels were significantly higher in dry eye patients than healthy people. A logistic regression model combining miR-450b-5p, miR-1283, miR-3671, and TGF-β2 was performed. This model presented a high discriminating value (AUC: 0.907, 0.876-0.939, P<0.001) than any single indicator, and the sensitivity and specificity were 77.7% and 92.7%, respectively. Compared with the low miR-450b-5p, low miR-1283, low miR-3671 and high TGF-β2 groups, the high miR-450b-5p, high miR-1283, high miR-3671 and low TGF-β2 groups had a significantly higher probability of corneal injury (TGF-β2: χ2=5.762, P=0.016; miR-450b-5p: χ2=13.267, P<0.001; miR-1283: χ2=19.431, P<0.001; miR-3671: χ2=8.131, P=0.004). Conclusion: Current model combining tear miR-450b-5p, miR-1283, miR-3671, and TGF-β2 had important values in the identification of dry eye and was of great value in evaluating the risk of corneal injury.

背景：迄今为止，尚未报道干眼患者和健康人的眼泪中microRNA（miR）表达谱的差异。在当前研究中，我们评估了miRs和转化生长因子beta2（TGF-β2）在区分干眼症中的意义。方法：选择2017年10月至2018年10月的138例干眼症患者。同期收集了138名健康人。通过门诊，电话或病历对所有患者进行了12个月的随访，并记录了角膜损伤的时间。结果：与健康人相比，干眼患者的TGF-β2水平明显降低（P <0.05）。阵列分析，预测软件和双荧光素酶报告基因检测表明miR-450b-5p，miR-1283和miR-3671可以靶向TGF-β2表达。撕裂miR-450b-5p，miR-1283，干眼患者中miR-3671和miR-3671的水平显着高于健康人。进行了结合miR-450b-5p，miR-1283，miR-3671和TGF-β2的逻辑回归模型。该模型具有比任何单一指标更高的鉴别值（AUC：0.907、0.876-0.939，P <0.001），并且敏感性和特异性分别为77.7％和92.7％。与低miR-450b-5p，低miR-1283，低miR-3671和高TGF-β2组相比，高miR-450b-5p，高miR-1283，高miR-3671和低TGF-β2组具有角膜损伤的可能性更高（TGF-β2：χ2= 5.762，P = 0.016; miR-450b-5p：χ2= 13.267，P <0.001; miR-1283：χ2= 19.431，P <0.001; miR-3671： χ2= 8.131，P = 0.004）。结论：目前的模型结合了泪液miR-450b-5p，miR-1283，miR-3671，